AI Audit Trail & Decision Evidence | Dweve Trace

Signed, content-addressed, replayable evidence for AI decisions, model outputs, agent actions, and sourced facts. Review what happened under EU controls.

What is a Dweve trace?

A Dweve trace is a bounded record of an AI-assisted operation. It keeps the relevant inputs, versions, authority, actions and evidence together so another person can check what happened and what the record does not prove.

  • A log records events, an explanation helps a person understand a result, and a trace binds the decision context to a declared replay or verification scope.
  • Integrity of a record does not establish that a source was correct, a policy was fair or an external side effect happened.
  • The stack uses scoped records rather than one universal proof artifact. The responsible product or foundation owns the boundary of each claim.

Choose the audience that matches your question

The page contains three selectable readings of the same subject.

For consumers

An AI answer is only one part of the story. In observability terms, logs record events; a trace follows one request or case through its spans. Dweve's trace keeps the relevant record of what happened, which material and rules mattered, who had authority, what can be checked, and what still needs human judgement.

For businesses

A trace turns an AI-assisted operation into a record that can move between operations, risk, audit and the affected person. It preserves the decision context without pretending that integrity, evidence and accountability are the same claim.

For engineers

Trace is the stack-wide guide to evidence, provenance and replay contracts. It separates event capture, explanation, source lineage, integrity checks and re-execution, then shows where each Dweve system contributes a scoped record rather than one universal proof artefact.